Improved Testing of Soldered Busbar Interconnects on Silicon Solar Cells

Transcription

1 Improved Testing of Soldered usbar Interconnects on Silicon Solar Cells R. Klengel*, M. Petzold*, D. Schade**,. Sykes** *Fraunhofer Institute of Mechanics of Materials IWM (Halle, Germany) **XYZTEC b.v. (Panningen, The Netherlands) bstract The expected life of a PV solar cell is years. To ensure this the quality and reliability of the material and process parameters of solder joints that connect single cells into strings has to be assured. standardized test method or clearly defined good-bad-criteria are not available for PV products. Thus a product, manufacturer or joining technology comparison is not possible. Due to the lack of standards and commercial available test equipment PV module manufacturers often implement their own test solutions and specifications. However, these mostly have limited scientific proof, are difficult to interpret and lack optimisation. ecause there is no standardization these solutions cannot be used for comparative characterization of different manufacturers, cell types, base- and joining materials or joining technologies. The most commonly used test for solar cell ribbon interconnects is the Pull test, as used in micro-electronics. For thin, brittle, large area silicon solar cells substantial modifications of the test equipment and methodology are required to realize a suitable metrology. In addition the factors influencing the defect conditions in multi crystalline material, high strains around the solder connect, inhomogeneous contact interfaces has to be considered. Results show the development, optimization and evaluation of a test method for soldered bus bar interconnects on silicon a solar cell that is independent of product, manufacturer and cell assembly. comparison between this new method, its results and the previously mentioned issues is made. The test method demonstrates how cell breakage during a test can be avoided by optimising how it is mounted and guided. Key words: Photovoltaic, Solar Cell, Mechanical Testing, Quality Improvement, Solder Joint Characterization Motivation Reductions in the cost and performance of solar cell and module production are an ongoing requirement in the PV industry. There are two factors of essential importance. On the one hand the optimization of material properties with the goal to reach a high and long-term stabile efficiency. On the other hand reducing production costs, whilst ensuring the highest quality and reliability possible are required. These requirements place a high demand on test and inspection that enable investigation into the mechanical, electrical and micro-structural properties of a solar cell and how they are affected by the materials and processes used during its manufacture. Often basic methods are adopted from other industries like micro electronic or electronic packaging. ut these were modified, enhanced and optimized for appliance in the PV sector. asics Silicon solar cells are made from wafer disks with a thickness of 200 microns or less. They are manufactured by numerous etching, coating, screen printing, and firing steps. metallization grid on the front side collects the light generated charge and is connected to bus bars. Flat copper ribbons, soldered to the bus bars, transfer the current from one cell to another in a series connection to so called strings. Several strings sealed in a lamination process create a module. [1] Fig. 1: - Overview of a three cell string connected by two soldered copper ribbons each - scheme of the solar cell series connection by soldered copper ribbons The soldering process is a well known interconnection technology in the electronic packaging industry. Low costs and high reliability are the main advantages of this type of bond. For quality control, there are many known tests and standards. For example, the pull test for electronic components is specified in DIN EN [2]. The challenge is whether these test conditions can simply be transferred to solar cell soldering. There are

2 differences between soldering of under bump metallization (UM) and the under interconnector metallization (UIM) used in solar cell construction. Usually the size of an UM is at most only a few square millimetres. The width of an UIM is between 1 and 3 mm with lengths up to 156mm. Therefore, the current density in an UM is much higher than in an UIM. Voids for example are less problematic than thermo mechanical stress. unsoldered areas are clearly distinguishable from each other. Fig. 2: Solar cell after pull test with common concept due large area silicon disruptions no conclusion to the interconnection quality can be made The different coefficient of thermal expansion between copper and silicon results in mechanical stress after the soldering process. In some cases, this leads to interruption of the thin grid finger at the front side of the cell [3]. Charge carriers, occurring around the interrupted grid finger, are not able to dissipate. The efficiency of the harmed cell and subsequently the efficiency of the whole module are reduced. Pulling ribbons from pre-damaged cells results in large area silicon tears. Micro cracks underneath the UIM are forced to spread along silicon grain boundaries and cause large area silicon disruptions. Instead of the UIM, the silicon wafer is tested [1]. Figure 2 shows a silicon solar cell after pull test using a common test concept. One aim of the presented development is to restrain the spreading micro cracks during the pull test and avoiding the large area silicon tears what is indispensable for getting evaluable and reproducible results. nother important motivation is to provide test equipment which allows comparing the results indendent of cell geometry, condition of contact materials, soldering technology or manufacturer. Further goals were making the test easier to handle, reaching a higher throughput and getting clearer criteria to distinguish good or poor solder joints. Principle of the Test Method There are three main types of failure modes occurring through the pull test (Figure 3). The first and favoured is that the interconnection fails within the UIM. The pull forces between soldered and C Fig. 3: Failure modes after pull test - no wetting, no solder joint formation - failing within the UIM, good quality C- silicon disruptions The second is that no wetting took place. The UIM remains unsoldered and there are nearly no measurably pull forces. The third and most inappropriate one is that the silicon wafer underneath the UIM fails, causing large area silicon disruptions. When silicon has brittle fracture behaviour, nearly no pull forces are measured [1]. djustments in soldering conditions like low temperature and cooling rate may reduce this problem [4]. C Fig. 4: Overview of some test concepts +- cell mounting on an incline plane and stationary ribbon clamping lead to inconstant pull angles C- flexible horizontal cell mounting and no static deviating point for the ribbon lead to overshooting and inconstant pull angles D- inconstant bending stiffness of the ribbon lead to varying pull angles, for 180 deviating a very long ribbon is needed which is not available by taking sample from a string The pull test of solder joints of electronic components is standardized in DIN EN 61189, including the pull angle of 90 and a velocity of 200 mm/min. Nevertheless, there are many different D

3 concepts in use. Figure 4 gives a view over some solutions. Looking at the natural load case, the main direction of applied force is longitudinal to the solder ribbon. For that reason, a pull angle of 180 may seem to be more realistic. ut it is not the task of the pull test to simulate a realistic load case. The pull test is used as process control for the soldering process as well as for characterisation of the solder joint stability as a quality criterion. Figure 5 images one advantage of a 90 pull angle. 2x ribbon length Start End 1x ribbon length Start End Fig. 5: Scheme shows one of the advantages of performing test with 90 pull angle - for 180 pull angle a long ribbons and place of the double cell lengths are needed - for 90 pull angle only the ribbon and place of the cell length are needed Every micro crack can lead to failure of the silicon and afterwards to a brittle crack extension underneath the UIM. To restrain this crack extension, it was found that an opposing force applied at the point of brake prevented crack propagation. Single micro cracks will only lead to small silicon tears. Large area silicon tears will only occur if the cell has been badly pre-damaged. Development of Test Equipment and Routine simple way to combine these requirements with an innovative measurement strategy is to upgrade a common measuring unit which has a flexible data base and onboard visual inspection. The XYZTEC Condor 250 multifunctional test machine was selected. The machine includes a vertical moving z-axis equipped with a force sensor and an x-y-stage for horizontal movement of the test piece. ccording to DIN 61189, the pull angle should be 90. The X-Y table moves the cell at the same velocity so that the point of breakage remains under the measuring head. So the measuring range equates to the ribbon length. The pull test with 180 causes a measuring range twice as long as the ribbon (Figure 5). Due the point of breakage remains in the same position the opposing force can be easily applied by using a triangular anvil with rounded edges made from special low-wear material. Pressing the anvil onto the cell may cause additional damage. Therefore the cell has to be pulled to the edge of the anvil and should not be fixed at the table in vertical direction. Figure 6 shows a scheme of the developed anvil and a ribbon being pulled. Figure 7 gives a picture of the complete machine and a detail of test situation. Peel tweezers Ribbon Peel force Opposing force Opposing force anvil Silicon Fig. 6: Scheme of the function principle developed for the new solar cell pull test equipment Since solar cells and module manufacturing is mass production, there are special requirements for measuring equipment. The main requirements for the ribbon pull test are stability, simplicity and high throughput. dditionally, it is necessary to generate reports that allow easy interpretation of the results. Calculated mean values, measured curves, and pictures of the fracture interface should be included. Fig. 7: Image of the universal test machine equipped with the new developed solar cell work holder for solder ribbon pull test

4 The cell is placed in a cassette that can be removed from the machine. Using two cassettes, one can be in test while the other is being reloaded. The following test procedure was developed: 1) Inserting the cell in the work holder 2) Fixing the work holder on the machine 3) Move to start position of first ribbon 4) Clamp the ribbon in the pull tweezers 5) Starting the test of the first usbar 6) fter finish first test unclamp the ribbon 7) Move to the second start position 8) Repeat procedure from point 4) fter finishing the last ribbon, the work holder is moved automatically to the load position. For security demands, the whole work holder is pneumatically lowered while the table moves toward the next starting- or load position. The test of one cell with three bus bars takes less than five minutes. The user has to clamp and unclamp the ribbon and to change the solar cells. While the test is ongoing, the next cell can be loaded to the second cassette. fter finishing the test, the cassettes will be interchanged and the next test can be started immediately. It is possible to test any kind of silicon solar cell. The work holder with the cell cassette as well as the test routine can be adapted to every cell geometry and condition like two or three bus bars, cell thickness, ribbon dimensions and soldering method. While the test is ongoing, a camera looks at both the cell and the pulled ribbon. So the failure mode can be observed and easily documented. How the developed test machine looks like is shown in Figure 7. Results of Test Development Fig. 8: Front side bus bars after pull test performed with the new developed test equipment no silicon tears and clear distinguishable failing interface levels Silicon solar cells were tested from several manufacturers, in varying conditions and different geometries. For all cells a pull test velocity of at least 5mm per second could be performed without any significant silicon disruptions. Figure 8 shows three front side bus bars after pull test without any damaged. Due to the semi-automatic application flow the test is considerable accelerated and the handling is remarkable easier. The dependence of the test results on secondary conditions like influences of varying friction depending on test velocity, influences of different ribbon material expansion, influences of solder material etc. were extensive investigated and will be published in near future. Micro structural Correlation Fig. 9: Graph of pull test measurement (force over distance) with imaging of the corresponding failure interface - front side bus bar - back side bus bar The graphical analysis of the measurement results and the images of the fracture interface allow correlating the pull force values with the corresponding failure mode. Figure 9 shows measurement graphs of a front side bus bar and a back side bus bar (both connected with soldered and unsoldered areas alternating) with the images of the correlating fracture interface. Fig. 10: Detailed graph (force over distance) of pull test measurement on a front side bus bar

5 In the detail of a measurement graph, pictured in Figure 10, it is visible that every single change in the fracture interface is reproducible in the progression of the curve. That indicates it seems to be possible to conclude to the solder joint condition by interpreting the measure curve. ut to ensure this hypothesis and to give defined criteria for the curve interpretation, a lot more investigations are necessary. Subject of actual research is the possibility to correlate the pull test measurement curves and fracture interfaces with concrete interface weakening like voids, non wetted areas or micro cracks etc. before performing the pull test. Therefore different non-distructive analyzes of soldered bus bar interconnects like scanning acoustic microscopy (SM), X-ray inspection or electroluminescence were performed before the pull test. fterwards the results of the preliminary investigation have to be correlated with the measurement curves and the images of the fracture interface. Figure 11 shows the solder joint interface of bus bars investigated with non distructive methods. Fig. 11: Overview of used micro structural investigation methods for solder joint quality analyzes: - Scanning coustic Microscopy (SM) - 2D-xray inspection Conclusions Improving measurement quality for testing solder joints on silicon solar cells includes many aspects. Firstly there is no standardization. Different test conditions like pull angle and test velocity implicates difficulties concerning comparability between different users. The second aspect is that it has to be assured that only the UIM is tested instead of silicon underneath. The disadvantage of common testers is that the cell is fixed while the ribbon is pulled. This can be improved by applying an opposing force on the ribbon at the point of test load application. Silicon disruptions will then not lead to large area tears, but only when extensive pre-cracks are present over the solder joint. [1] With the developed test equipment and measurement routine it is succeeded to provide a reproducible, fast and easy handable method which is usable indendent from solar cell geometry, condition of contact materials, soldering technology or manufacturer. It is succeeded to show that it is possible to correlate the measurement curve of the pull test with the fracture interface. This allows to interprete the results with statements in terms of the solder joint quality. To ensure these interpretation hypothesis further investigations were performed in present time using non-distructive test methods before the pull test. The sum of results from preliminary investigations, pull test measurement and analysis of fracture interface should allow defining criteria for rating the solder joint quality. cknowledgements This work was partly supported by the German Ministry of Economics and Technology under contract TSOLOT (FKZ01FS10010). Many thanks to Tino Stephan, Marcel Mittag and René Härtel for the great assistance in this work. References [1] J. Wendt, R. Klengel, D. Schade: Improved quality test method for solder ribbon interconnects on silicon solar cells ; 12th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), June 2010 Las Vegas, Nevada US [2] DIN EN 61189: Test methods for electrical materials, printed boards and other interconnection structures and assemblies. German DIN standard [3] J. Wendt, M. Träger, M. Mette: The link between mechanical stress induced by soldering and micro damages in silicon solar cells ; 24th European Photovoltaic Solar Energy Conference, September 2009, Hamburg, Germany [4].M. Gabor: Soldering Induced Damage To Thin Si Solar Cells nd Detection Of Cracked Cells In Modules ; 21st European Photovoltaic Solar Energy Conference, 4-8 September 2006, Dresden, Germany